Well logging lithology identification method based on cuckoo search clustering analysis

Abstract

The invention discloses a logging lithology identification method based on cuckoo search clustering analysis, and belongs to the technical field of lithology identification in logging information. The method comprises the following steps: 1) selecting logging curve data; 2) standardization processing of logging curve data; the method comprises the following steps of (1) standardizing data, (2) standardizing the data, (3) carrying out clustering analysis on the data after standardization processing based on a cuckoo search clustering analysis algorithm, endowing various types of lithology and obtaining a lithology dividing boundary, and (4) inputting the obtained lithology dividing boundary into drawing software and carrying out lithology recognizing.The formation lithology of a research area can be quickly, accurately and efficiently recognized, and an effective theoretical support is provided for logging evaluation.

Description

Technical Field

[0001] The invention relates to a lithology identification method, in particular to a well logging lithology identification method based on cuckoo search cluster analysis, and belongs to the technical field of lithology identification in well logging data. Background Art

[0002] Lithology feature identification is the basis for understanding formations and reservoir prediction. Accurate identification of lithology is of great significance to reservoir division, oil and gas layer division and reservoir evaluation. Compared with the method of identifying lithology through coring and lithology scanning, the use of logging data to identify lithology has the characteristics of fast speed and low cost, and is widely used.

[0003] At present, the main methods for identifying formation lithology using logging data are: intersection diagram method (Study on well logging identification method of sandstone conglomerate lithofacies, Fu Jianwei et al.; Characteristics of conglomerate layer of Shawan Formation in Chunfeng Oilfield and quantitative identification by well logging, Hou Qingjie et al.), that is, selecting two lithology-sensitive logging curves to make an intersection diagram to determine the boundary line of different lithologies. However, for formations with complex lithologies, different lithology points overlap on the intersection diagram and are difficult to distinguish. For example, the prior art CN111458767A uses the intersection diagram method to identify lithology; The lithology identification method based on machine learning (Research and application of logging lithology identification method based on improved multi-classification twin support vector machine, Su Fu et al.; Application of GA-BP neural network model in logging lithology identification in Binchang mining area, Liu Mingjun et al.; Neural network lithology identification method based on principal component analysis and learning vectorization, Hu Hong et al.) is mainly: by selecting logging data of different typical lithologies, a data set is established, and then the model is trained based on this training set, and the trained model is used for lithology identification. This method can quickly identify lithology, but its accuracy depends on the quality of the training set; The lithology identification method based on cluster analysis (using Matlab cluster analysis toolbox to divide carbonate rock lithology, Zhong Yajun et al.; lithology identification research based on fuzzy C-means clustering, Li Haiyan et al.) mainly includes: performing principal component analysis on the logging curve, dividing the retained principal component data into multiple categories through cluster analysis, and the operator then assigns the actual lithology to each category to achieve the effect of lithology identification. This method is fast and accurate, but it does not take into account some shortcomings of the clustering algorithm itself, such as the clustering results rely on the initial cluster center and there is a local convergence situation, such as: the prior art CN109113729A lithology identification method and device based on logging curves.

[0004] Therefore, it is necessary to propose an efficient and accurate method for identifying lithology to meet the needs of well logging evaluation. Summary of the invention

[0005] The present invention aims to overcome the deficiencies of the prior art and provide a well logging lithology identification method based on cuckoo search cluster analysis, which is convenient for quickly, accurately and efficiently identifying the lithology of the formation in the study area and provides an effective theoretical basis for well logging evaluation.

[0006] In order to achieve the above technical objectives, the following technical solutions are proposed: The first purpose of this technical solution is to propose a well logging lithology identification method based on cuckoo search cluster analysis, comprising the following steps: S1: Selection of logging curve data Select the wells with logging, mud logging and lithology data in the work area, select the standard layers corresponding to different lithologies in the wells, and then use the characteristic curves in the corresponding standard layers that are sensitive to lithology (high lithology recognition) as implementation data (data for subsequent processing), and then obtain the logging curve data; Among them, the characteristic curves include: density curve, compensated neutron curve, sound wave curve and natural gamma curve; S2: Standardization / normalization of logging curve data Determine the standard well and standard layer in the existing well data, use the peak value method to calibrate the logging curves of all wells in the selected work area, and use the deviation standardization method to standardize the data obtained in step S1, that is, to make the mathematical units between the data consistent; S3: Perform cluster analysis on the data after the standardization process in step S2 based on the cuckoo search cluster analysis algorithm, assign lithology to each category, and obtain lithology boundaries, specifically: a. Initialize clustering parameters, including: the number of cluster centers (k), number of iterations, minimum step size, maximum step size, discovery probability, Levy index, and clustering accuracy; b. Randomly initialize k cluster centers; c. Use the following formula (1) to obtain the distance from each point to the cluster center; (1); in, For the data after standardization, For the location of the cluster center, divide each standardized data into the cluster center closest to it; d. Obtain the cluster center after division according to the mean of each cluster, and calculate the fitness of each cluster center according to the following formula (2); (2); in, For the The sum of the distances from all data points in a class to the cluster center, For the The number of all data points in each class; Then, the current optimal cluster center and its fitness value are retained, and the metric value of the current clustering result is obtained according to the following formula (3); (3); e. Keep the optimal cluster center of the previous iteration and obtain the control value of the step length according to the following formula (4); (4); in, is the minimum step length, is the maximum step length, is the number of iterations, is the maximum number of iterations; Then, according to the following formula (5) proposed by Mantegna, the update step size is obtained, that is, the remaining k-1 cluster centers are updated; (5); in, obey Normal distribution, obey Normal distribution, and , , , is a stable parameter; f. Re-divide the remaining logging data into the nearest cluster center, obtain the cluster center of the new class according to the mean, and obtain the new fitness value and cluster result metric value, and compare this metric value with the previous generation result: if the metric value of the current generation cluster result is greater than the metric value of the previous generation cluster result, then update the combination of cluster centers; otherwise, do not update the combination of cluster centers; g. When the number of iterations is equal to the maximum number of iterations, or the measurement result meets the accuracy requirement, the process ends; otherwise, a random number is generated for each cluster center other than the optimal cluster center, and compared with the discovery probability. If the random number is less than the discovery probability, the cluster center is abandoned, a new cluster center is randomly generated, and then step eg is repeated; h. Define the category to which the logging data belongs, assign lithology to each category of logging data according to the characteristics of the logging curve, and obtain the lithology boundary; S4: Input the lithology boundary obtained in step S3 into a drawing software (such as Resform software, ciflog software or techlog software) to perform lithology identification.

[0007] The second objective of the technical solution is to propose a well logging lithology identification system based on cuckoo search cluster analysis, characterized in that it includes: The well logging curve data selection module is used to select wells with logging, mud logging and lithology data in the work area, select standard layers corresponding to different lithologies in the wells, and then use the characteristic curves in the corresponding standard layers that are sensitive to lithology as implementation data to obtain well logging curve data; The logging curve data standardization processing module is used to determine the standard wells and standard layers in the existing well data, use the peak value method to correct the logging curves of all wells in the selected work area, and use the deviation standardization method to standardize the data obtained by the logging curve data selection module; The cluster analysis module, based on the cuckoo search cluster analysis algorithm, performs cluster analysis on the standardized data obtained by the logging curve data standardization processing module, assigns lithology to each category, and obtains lithology boundaries; Lithology judgment module: input the lithology boundaries obtained by the cluster analysis module to complete the identification of lithology judgment.

[0008] Preferably, in the cluster analysis module, the cuckoo search cluster analysis algorithm includes: a. Initialize clustering parameters, including: the number of cluster centers (k), number of iterations, minimum step size, maximum step size, discovery probability, Levy index, and clustering accuracy; b. Initialize k cluster centers; c. Use the following formula (1) to obtain the distance from each point to the cluster center; (1); in, For the data after standardization, For the location of the cluster center, divide each standardized data into the cluster center closest to it; d. Obtain the cluster center after division according to the mean of each cluster, and calculate the fitness of each cluster center according to the following formula (2); (2); in, For the The sum of the distances from all data points in a class to the cluster center, For the The number of all data points in each class; Then, the current optimal cluster center and its fitness value are retained, and the metric value of the current clustering result is obtained according to the following formula (3); (3); e. Keep the optimal cluster center of the previous iteration and obtain the control value of the step length according to the following formula (4); (4); in, is the minimum step length, is the maximum step length, is the number of iterations, is the maximum number of iterations; Then, according to the following formula (5) proposed by Mantegna, the update step size is obtained, that is, the remaining k-1 cluster centers are updated; (5); in, obey Normal distribution, obey Normal distribution, and , , is a stable parameter; f. Re-divide the remaining logging data into the nearest cluster center, obtain the cluster center of the new class according to the mean, and obtain the new fitness value and cluster result metric value, and compare this metric value with the previous generation result: if the metric value of the current generation cluster result is greater than the metric value of the previous generation cluster result, then update the combination of cluster centers; otherwise, do not update the combination of cluster centers; g. When the number of iterations is equal to the maximum number of iterations, or the measurement result meets the accuracy requirement, the process ends; otherwise, a random number is generated for each cluster center other than the optimal cluster center, and compared with the discovery probability. If the random number is less than the discovery probability, the cluster center is abandoned, a new cluster center is randomly generated, and then step eg is repeated; h. Define the category to which the logging data belongs, assign lithology to each category of logging data according to the characteristics of the logging curve, and obtain the lithology boundary.

[0009] The third objective of the present technical solution is to provide: a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps of any one of the methods described in the above technical solutions.

[0010] The fourth objective of the present technical solution is to provide: an information data processing terminal for a well logging lithology identification method based on cuckoo search cluster analysis.

[0011] The beneficial technical effects brought about by adopting this technical solution are: The present invention selects multiple curves that are sensitive to lithology for clustering and uses the cuckoo optimization clustering analysis algorithm to effectively avoid the problem of local convergence of clustering results and the dependence of results on initial values. The clustering accuracy is improved, and then the lithology of each category is assigned according to the characteristics of the lithology curve, and the lithology boundary line is obtained. The obtained lithology boundary line has high accuracy, and then the lithology of the target well is identified by this lithology boundary line; The present invention can combine a variety of methods to make lithology identification more accurate. For example, in the logging curve selection stage, principal component analysis can be used to reduce the dimension, and then this method can be used to perform cluster analysis, and finally the lithology corresponding to each classification result can be assigned; or this method can be used to preliminarily classify the lithology, and then discriminant analysis can be used to subdivide the small categories, etc.; for formations with complex lithology, two or three logging curves may be difficult to distinguish the lithology of the formation, and multiple logging curves can be selected for cluster analysis. Multiple curves can better utilize the formation information, so that the lithology identification result is better. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] Figure 1 This is a flow chart of the improved clustering algorithm involved in Example 1; Figure 2 This is a flow chart of lithology identification involved in Example 1; Figure 3 This is the lithology identification effect in Example 5; Figure 4 This is the clustering result in Example 6. DETAILED DESCRIPTION

[0013] The following is a clear and complete description of the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0014] Example 1 This embodiment provides a well logging lithology identification method based on cuckoo search cluster analysis, such as Figure 2 , including the following steps: S1: Selection of logging curve data Select the wells with logging, mud logging and lithology data in the work area, select the standard layers corresponding to different lithologies in the wells, and then use the characteristic curves in the corresponding standard layers that are sensitive to lithology (high lithology recognition) as implementation data (data for subsequent processing), and then obtain the logging curve data; Among them, the characteristic curves include: density curve, compensated neutron curve, sound wave curve and natural gamma curve; S2: Standardization / normalization of logging curve data Determine the standard well and standard layer in the existing well data, use the peak value method to calibrate the logging curves of all wells in the selected work area, and use the deviation standardization method to standardize the data obtained in step S1, that is, to make the mathematical units between the data consistent; S3: Perform cluster analysis on the data after the standardization process in step S2 based on the cuckoo search cluster analysis algorithm, assign lithology to each category, and obtain lithology boundaries, specifically: a. Initialize clustering parameters, including: the number of cluster centers (k), number of iterations, minimum step size, maximum step size, discovery probability, Levy index, and clustering accuracy; b. Randomly initialize k cluster centers; c. Use the following formula (1) to obtain the distance from each point to the cluster center; (1); in, For the data after standardization, For the location of the cluster center, divide each standardized data into the cluster center closest to it; d. Obtain the cluster center after division according to the mean of each cluster, and calculate the fitness of each cluster center according to the following formula (2); (2); in, For the The sum of the distances from all data points in a class to the cluster center, For the The number of all data points in each class; Then, the current optimal cluster center and its fitness value are retained, and the metric value of the current clustering result is obtained according to the following formula (3); (3); e. Keep the optimal cluster center of the previous iteration and obtain the control value of the step length according to the following formula (4); (4); in, is the minimum step length, is the maximum step length, is the number of iterations, is the maximum number of iterations; According to the following formula (5) proposed by Mantegna, Get the update step size, that is, update the remaining k-1 cluster centers; (5); in, obey Normal distribution, obey Normal distribution, and , , , is a stable parameter; f. Re-divide the remaining logging data into the nearest cluster center, obtain the cluster center of the new class according to the mean, and obtain the new fitness value and cluster result metric value, and compare this metric value with the previous generation result: if the metric value of the current generation cluster result is greater than the metric value of the previous generation cluster result, then update the combination of cluster centers; otherwise, do not update the combination of cluster centers; g. When the number of iterations is equal to the maximum number of iterations, or the measurement result meets the accuracy requirement, the process ends; otherwise, a random number is generated for each cluster center other than the optimal cluster center, and compared with the discovery probability. If the random number is less than the discovery probability, the cluster center is abandoned, a new cluster center is randomly generated, and then step eg is repeated; h. Define the category to which the logging data belongs, assign lithology to each category of logging data according to the characteristics of the logging curve, and obtain the lithology boundary; S4: Input the lithology boundary obtained in step S3 into a drawing software (such as Resform software, ciflog software or techlog software) to perform lithology identification.

[0015] This embodiment can combine multiple methods to make lithology identification more accurate. For example, in the logging curve selection stage, principal component analysis can be used to reduce the dimension, and then this method can be used to perform cluster analysis, and finally the lithology corresponding to each classification result can be assigned; or this method can be used to preliminarily classify the lithology, and then discriminant analysis can be used to subdivide the small categories, etc.; for formations with complex lithology, two or three logging curves may be difficult to distinguish the lithology of the formation, and multiple logging curves can be selected for cluster analysis. Multiple curves can better utilize the formation information, so that the lithology identification result is better.

[0016] In addition, in order to improve the accuracy of lithology identification based on cluster analysis, this embodiment uses clustering algorithms (such as Figure 1 ) has been adaptively improved, that is, the cuckoo search clustering analysis algorithm mentioned above (such as Figure 2 ).

[0017] Example 2 This embodiment provides a well logging lithology identification system based on cuckoo search cluster analysis, including: The well logging curve data selection module is used to select wells with logging, mud logging and lithology data in the work area, select standard layers corresponding to different lithologies in the wells, and then use the characteristic curves in the corresponding standard layers that are sensitive to lithology as implementation data to obtain well logging curve data; The logging curve data standardization processing module is used to determine the standard wells and standard layers in the existing well data, use the peak value method to correct the logging curves of all wells in the selected work area, and use the deviation standardization method to standardize the data obtained by the logging curve data selection module; The cluster analysis module, based on the cuckoo search cluster analysis algorithm, performs cluster analysis on the standardized data obtained by the logging curve data standardization processing module, assigns lithology to each category, and obtains lithology boundaries; Lithology judgment module: input the lithology boundaries obtained by the cluster analysis module to complete the identification of lithology judgment.

[0018] Wherein, in the cluster analysis module, the cuckoo search cluster analysis algorithm includes: a. Initialize clustering parameters, including: the number of cluster centers (k), number of iterations, minimum step size, maximum step size, discovery probability, Levy index, and clustering accuracy; b. Randomly initialize k cluster centers; c. Use the following formula (1) to obtain the distance from each point to the cluster center; (1); in, For the data after standardization, For the location of the cluster center, divide each standardized data into the cluster center closest to it; d. Obtain the cluster center after division according to the mean of each cluster, and calculate the fitness of each cluster center according to the following formula (2); (2); in, For the The sum of the distances from all data points in a class to the cluster center, For the The number of all data points in each class; Then, the current optimal cluster center and its fitness value are retained, and the metric value of the current clustering result is obtained according to the following formula (3); (3); e. Keep the optimal cluster center of the previous iteration and obtain the control value of the step length according to the following formula (4); (4); in, is the minimum step length, is the maximum step length, is the number of iterations, is the maximum number of iterations; According to the following formula (5) proposed by Mantegna, Get the update step size, that is, update the remaining k-1 cluster centers; (5); in, obey Normal distribution, obey Normal distribution, and , , , is a stable parameter; f. Re-divide the remaining logging data into the nearest cluster center, obtain the cluster center of the new class according to the mean, and obtain the new fitness value and cluster result metric value, and compare this metric value with the previous generation result: if the metric value of the current generation cluster result is greater than the metric value of the previous generation cluster result, then update the combination of cluster centers; otherwise, do not update the combination of cluster centers; g. When the number of iterations is equal to the maximum number of iterations, or the measurement result meets the accuracy requirement, the process ends; otherwise, a random number is generated for each cluster center other than the optimal cluster center, and compared with the discovery probability. If the random number is less than the discovery probability, the cluster center is abandoned, a new cluster center is randomly generated, and then step eg is repeated; h. Define the category to which the logging data belongs, assign lithology to each category of logging data according to the characteristics of the logging curve, and obtain the lithology boundary.

[0019] Example 3 Based on Example 1, this embodiment provides: a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the steps of any one of the methods described in the technical solution of Example 1 are implemented.

[0020] Example 4 Based on the first embodiment, the present embodiment provides: an information data processing terminal for a well logging lithology identification method based on cuckoo search cluster analysis.

[0021] Example 5 Based on Example 1, this example takes the well logging curve of a certain well as an example, selects three logging curves of natural gamma, compensated density and compensated neutron for lithology identification, clusters and separates the natural gamma value of mudstone greater than 90API, and the natural gamma value of dolomite less than 90API, and then selects density and neutron logging values ​​for clustering, assigning each category of lithology, and the density of sand-crushed dolomite is less than 2.74g / cm 3 , neutron is 6.5-7.8%, neutron of gray dolomite is less than 6.5%, density is 2.74-2.77g / cm3 , the neutrons of pink dolomite are 7.8-11.5%, and the density is 2.77-2.85g / cm 3 , input this lithology boundary into Resform software, and the final lithology identification result is as follows Figure 3 shown.

[0022] It can be seen that the present invention provides an algorithm with more accurate clustering results, and the boundaries obtained by applying this method can be used for lithology identification with high accuracy.

[0023] Example 6 Based on Example 1, this example adopts an improved method - cuckoo search cluster analysis algorithm to conduct experiments, selects compensation density curve and compensation neutron curve values ​​for experiments, first removes outliers in the data, then normalizes the data to eliminate the impact of dimension, and finally uses the preprocessed data as implementation data to perform cluster analysis using the improved method.

[0024] The number of cluster centers is set to 3, i.e., 3 types of lithology. The range of the clustering results corresponding to each category is ( Figure 4 ): Class I DEN<2.6g / cm 3 ; Type II DEN>2.6g / cm 3 , CNL<7%; Class III DEN>2.6g / cm 3 , CNL>7%. According to the conventional experience of the staff in this field, the lithology density is less than 2.6g / cm 3 , and then the rock density is greater than 2.6g / cm 3 Classification based on the number of neutrons, that is, classification based on the content of cloud content, is in line with the actual lithology.

Claims

1. A well logging lithology identification method based on cuckoo search cluster analysis, It is characterized in that The steps include: Selection of well logging data; Standardized processing of well logging curve data; Based on the cuckoo search cluster analysis algorithm, cluster analysis is performed on the standardized data, and lithology is assigned to each category to obtain the lithology boundaries; The obtained lithology boundaries are input into the drawing software for lithology identification.

2. The well logging lithology identification method based on cuckoo search cluster analysis according to claim 1, It is characterized in that The selection of logging curve data includes: selecting wells with logging, mud logging and lithology data in the work area, selecting standard layers corresponding to different lithologies in the wells, and then using characteristic curves in the corresponding standard layers that are sensitive to lithology as logging curve data.

3. The well logging lithology identification method based on cuckoo search cluster analysis according to claim 2, It is characterized in that The characteristic curves include: a density curve, a compensated neutron curve, a sound wave curve and a natural gamma curve.

4. The well logging lithology identification method based on cuckoo search cluster analysis according to claim 3, It is characterized in that The standardized processing of the logging curve data includes: determining the standard wells and standard layers in the existing well data, correcting the logging curves of all wells in the selected work area using the peak method, and standardizing the logging curve data using the deviation standardization method.

5. The well logging lithology identification method based on cuckoo search cluster analysis according to claim 4, It is characterized in that Based on the cuckoo search cluster analysis algorithm, cluster analysis is performed on the standardized data, and each category of lithology is assigned to obtain the lithology boundaries, including: a. Initialize clustering parameters, including the number of cluster centers k, number of iterations, minimum step size, maximum step size, discovery probability, Levy index, and clustering accuracy; b. Randomly initialize cluster centers; c. Use the following formula (1) to obtain the distance from each point to the cluster center; （1）； in, For the data after standardization, For the location of the cluster center, divide each standardized data into the cluster center closest to it; d. Obtain the cluster center after division according to the mean of each cluster, and obtain the fitness of each cluster center according to the following formula (2); （2）； in, For the The sum of the distances from all data points in a class to the cluster center, For the The number of all data points in each class; Then, the current optimal cluster center and its fitness value are retained, and the metric value of the current clustering result is obtained according to the following formula (3); （3）； e. Keep the optimal cluster center of the previous iteration and obtain the control value of the step length according to the following formula (4); （4）； in, is the minimum step length, is the maximum step length, is the number of iterations, is the maximum number of iterations; Then, according to the following formula (5) proposed by Mantegna, the update step size is obtained, that is, the remaining k-1 cluster centers are updated; （5）； in, obey Normal distribution, obey Normal distribution, and , , , is a stable parameter; f. Re-divide the remaining logging data into the nearest cluster center, obtain the cluster center of the new class according to the mean, and obtain the new fitness value and cluster result metric value, and compare this metric value with the previous generation result: if the metric value of the current generation cluster result is greater than the metric value of the previous generation cluster result, then update the combination of cluster centers; otherwise, do not update the combination of cluster centers; g. When the number of iterations is equal to the maximum number of iterations, or the measurement result meets the accuracy requirement, the process ends; otherwise, a random number is generated for each cluster center other than the optimal cluster center, and compared with the discovery probability. If the random number is less than the discovery probability, the cluster center is abandoned, a new cluster center is randomly generated, and then step eg is repeated; h. Define the category to which the logging data belongs, assign lithology to each category of logging data according to the characteristics of the logging curve, and obtain the lithology boundary.

6. The well logging lithology identification method based on cuckoo search cluster analysis according to claim 5, It is characterized in that The drawing software is Resform software, ciflog software or techlog software.

7. A well logging lithology identification system based on cuckoo search cluster analysis, It is characterized in that include: The well logging curve data selection module is used to select wells with logging, mud logging and lithology data in the work area, select standard layers corresponding to different lithologies in the wells, and then use the characteristic curves in the corresponding standard layers that are sensitive to lithology as implementation data to obtain well logging curve data; The logging curve data standardization processing module is used to determine the standard wells and standard layers in the existing well data, use the peak value method to correct the logging curves of all wells in the selected work area, and use the deviation standardization method to standardize the data obtained by the logging curve data selection module; The cluster analysis module, based on the cuckoo search cluster analysis algorithm, performs cluster analysis on the standardized data obtained by the logging curve data standardization processing module, assigns lithology to each category, and obtains the lithology boundaries; Lithology judgment module: input the lithology boundaries obtained by the cluster analysis module to complete the identification of lithology judgment.

8. The well logging lithology identification system based on cuckoo search cluster analysis according to claim 7, It is characterized in that In the cluster analysis module, the cuckoo search cluster analysis algorithm includes: a. Initialize clustering parameters, including: the number of cluster centers k, number of iterations, minimum step size, maximum step size, discovery probability, Levy index, and clustering accuracy; b. Randomly initialize k cluster centers; c. Use the following formula (1) to obtain the distance from each point to the cluster center; （1）； in, For the data after standardization, For the location of the cluster center, divide each standardized data into the cluster center closest to it; d. Obtain the cluster center after division according to the mean of each cluster, and calculate the fitness of each cluster center according to the following formula (2); （2）； in, For the The sum of the distances from all data points in a class to the cluster center, For the The number of all data points in each class; Then, the current optimal cluster center and its fitness value are retained, and the metric value of the current clustering result is obtained according to the following formula (3); （3）； e. Keep the optimal cluster center of the previous iteration and obtain the control value of the step length according to the following formula (4); （4）； in, is the minimum step length, is the maximum step length, is the number of iterations, is the maximum number of iterations; Then, according to the following formula (5) proposed by Mantegna, the update step size is obtained, that is, the remaining k-1 cluster centers are updated; （5）； in, obey Normal distribution, obey Normal distribution, and , , , is a stable parameter; f. Re-divide the remaining logging data into the nearest cluster center, obtain the cluster center of the new class according to the mean, and obtain the new fitness value and cluster result metric value, and compare this metric value with the previous generation result: if the metric value of the current generation cluster result is greater than the metric value of the previous generation cluster result, then update the combination of cluster centers; otherwise, do not update the combination of cluster centers; g. When the number of iterations is equal to the maximum number of iterations, or the measurement result meets the accuracy requirement, the process ends; otherwise, a random number is generated for each cluster center other than the optimal cluster center, and compared with the discovery probability. If the random number is less than the discovery probability, the cluster center is abandoned, a new cluster center is randomly generated, and then step eg is repeated; h. Define the category to which the logging data belongs, assign lithology to each category of logging data according to the characteristics of the logging curve, and obtain the lithology boundary.

9. A computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the well logging lithology identification method based on cuckoo search cluster analysis as described in any one of claims 1 to 6.

10. An information data processing terminal according to the well logging lithology identification method based on cuckoo search cluster analysis according to any one of claims 1 to 6.

Images